Relationship of Freestyle Swimming Performance Between Adolescence and Adult Japanese Top - class Swimmers

Objective : The purpose of this study was to evaluate the relationship of freestyle swimming performance between adolescence and adult Japanese top-class swimmers. Methods : The 234 males and 134 females ranked in the top 100 records of 50, 100, 200 and 400 m freestyle short-course performances of over 22 years old during 2016 to 2018 (latest TIME), and their season best performances during 2006 to 2017 were analyzed. All data were collected from the public database“Swimrecord.com”by Japan Swimming Federation. Performance maturity status was determined by the ratio of each age record to the latest TIME and the relationship between the latest TIME and each age record was evaluated. Also, the subjects were divided into the upper or lower group with reference to the latest TIME and the developmental changes of performance were compared between the groups. Results : The performance maturity status was higher in females than males at each age. The correlation between the latest TIME and each age record became gradually stronger from males aged 17 years and females aged 14 years, but there was no clear correlation in the previous records. There was no difference in developmental changes in performance between the upper and lower groups in both males and females. Conclusions : The longitudinal analysis of the swimming performance in the growth period reveals that the Japanese top-class swimmers show high performance before 12 years old and that the time to approximate peak performance appeared earlier in females than in males and the relationship is different depending on swimming distance.


Introduction
To define a relationship between the age and performance of athletes during the growing stage can be useful for measuring their future potential and selecting appropriate training options 1) . Generally speaking, the age when athletes achieve their best performance during the growing stage from childhood to adulthood and the course of time of improvements in their competitiveness are consid-ered to vary by the type of sports. In swimming, for example, it has been reported that athletes tend to show their best performance earlier than other sports 2) . However, recent studies report some cases where swimmers perform at their best in adulthood 3) , which suggests the need to re-evaluate the relationship between age and performance of competitive swimmers.
There are also several studies regarding the performance of top-class competitive swimmers during the growing stage which aims to not only discover and develop promising athletes but also determine achievable short-term goals and form long-term plans toward their target competition. Initiatives to help early discovery of promising young athletes include analyzing the variation pattern of the level of competitions where elite swimmers with experiences in the Olympics Games and other international competitions have participated during the growing stage, along with the relationship between their performance and other factors including their coach, training, and environment 4) 5) . Other countries have reported that the analysis on performance variations of top-class athletes during the growing stage by tracking back through their records stored in the database gives information necessary for properly selecting training methods and setting goals for promising young swimmers 6) 7) . Thus, conducting longitudinal analysis on the characteristics of performance variations of competitive swimmers, then utilizing the data in choosing proper training methods is expected to largely contribute to improvements in their competitiveness.
Recently, an increasing number of countries have been compiling databases of the records of the regional or world competitions. Japan also has a public database called"Swimrecord.com"that stores records of domestic and international competitions officially accredited by the Japan Swimming Federation. This public database organizes the collected records in a national ranking list to provide the domestic ranking of individual athletes, in addition to the updated official records of individual athletes in each competition available for athletes and coaches. Such databases are used for developing a performance forecasting model by combining the stored records and physical characteristic parameters, and for choosing proper training methods 8) 9) . It is also reported that the analysis of the relationship between age and performance during adolescence, and the robustness of the records can be useful information for coaches 10) . However, these reports only analyze the adolescence of male swimmers. It is not able to sufficiently explain the characteristics of swimmers who are supposed to show peak performance during adulthood. Also, to the best of our knowledge, no research has conducted a longitudinal analysis on the age-associated changes of Japanese competitive swimmersʼ performance by using the database published in Japan.
Therefore, this research conducted longitudinal analysis on the past records of top-class male and female adult swimmers who participated in short-course freestyle swimming competition by using the public database storing all records of official competitions accredited by the Japan Swimming Federation, through which it aimed to evaluate the relationship of freestyle swimming performance between adolescence and adulthood of Japanese top-class swimmers.

Methods
In order to track back through the records of freestyle short-course performances from adulthood to adolescence, the database website that contains the results of the Japan Swimming Federation competition and the results of the Japanese representative swimmers at overseas competitions were used ( "Swimrecord.com" ; http://www.swimrecord.com/index.html). "Swimrecord.com" contains the top 100 and the top 50 records of elementary schools, junior high schools, high schools and universities in a total of 53 categories including 47 prefectures in Japan and the 6 branches of university student federations in Japan. In addition, the top 100 rankings for all of these records and the top 50 national rankings for each class are shown. Thus, in order to analyse swimmers ranked top in the national ranking we analyzed the records ranked in the top 100 shown in this national ranking in this study. However, among the data included in this database, the data for 2005 and earlier are inconsistent, in particular because the data of primary school students are not stored sufficiently. Thus the research subjects in this research were the official record of freestyle shortcourse swimming recorded from April 2006 to March 2017 in this database, approximately 3.9 million.

2) Extraction of data for tracking back
The swimming records to track back the candidate for analysis from adulthood to adolescence (freestyle short-course eventʼs record in the prefectural swimming association of the whole country during April 2006 to March 2017) were extracted from the public database in text data format, then organized in the order of performance sorted by gender and swimming distance (50 m, 100 m, 200 m and 400 m).

3) Determination of subject for analysis
Both organized text data were verified retrospectively for each candidate of the analysis, and the records that could be extracted for a total of over six seasons were used for the analysis. The extracted records for over 22 years old swimmers (hereinafter:"latest TIME" , males: 234, females: 134) and their records of the period from 12 to 21 years old (hereinafter:"school-age TIME" , males: 1,775, females: 1,109) were organized by gender and swimming distance.

4) Anonymization and conversion to numerical data
The latest TIME and school-age TIME were anonymized and converted from minutes and seconds to seconds before analysis.

Maturity status
As an indicator to identify the characteristics of performance variations, the latest TIME and school-age TIME were converted in swimming speed, which allowed for calculating the ratio of swimming speed of the school-age TIME to that of the latest TIME (hereinafter"performance maturity" ).

Developmental changes in performance in the upper and lower groups
The top-class swimmers were divided into two groups; upper or lower level with reference to the median of the latest TIME of each distance in males and females. The developmental changes in performance were compared between the upper and lower groups.

Statistical Analysis
We evaluated the relationship between the latest and school-age TIMEs sorted by gender and swimming distance and each school-age TIME using Pearsonʼs product moment correlation coefficient was used to evaluate the relationship between the variables. In addition, Spearmanʼs rank correlation was used to check if P earsonʼs correlation coefficient was incorrectly estimated high when the number of data was small. Two-way ANOVA was used to examine the performance maturity (gender and age) and the developmental changes in performance (performance level and age). If a significant effect was found in the interaction, the test of the simple main effect at each level of each factor was examined, and a post hoc Bonferroni correction for multiple comparisons was used to identify significant differences. The p < 0.05 criterion was used for establishing statistical significance. All statistical procedures were performed using BellCurve for Excel (Social Survey Research Information Co., Ltd.).

Ethics
The use of data from"Swimrecord.com"for this research was approved by the Japan Swimming Federation. All data used in the analysis was anonymized before performing statistical analysis. This research was accredited by the ethics committee of Juntendo University Graduate School of Health and Sports Science (approved #29-141).

The performance maturity status
Table-1 and -2 show the school-age TIME of each of the swimming distances respectively for male and female swimmers. The average schoolage TIME of each of the swimming distances showed a tendency to shorten from adolescence to adulthood for males and females. The performance maturity status of males and females (ratio of school-age TIME to latest TIME) at the school-age was shown each swimming distance. The performance maturity status for 12 years old was 80.0 to 83.2% for males and 89.8 to 90.5% for females. Interactions related to maturity status were significant for gender and each school-age across all swim distances (Two-way ANOVA,  13.0; p < 0.05). Also, according to the results of multiple comparisons of the Maturity status in schoolage, females tended earlier than males when Maturity status approximated adult values (Figure-2). Table-3 and -4 shows the correlation coefficients between the latest TIME and school-age TIME by each distance and gender. The correlation coefficients Data are mean ± SD and number of subjects.

Figure-2
The performance maturity status of males and females (%; ratio of school-age TIME to latest TIME) at the school-age of each swimming distance of males at each school-age and each swimming distance (50 m, 100 m, 200 m and 400 m) were strong from 1 year to 3 or 5 years later, but gradually weakened after that. The correlation with the latest TIME became strong from 18 years old at 50 m, 17 years old at 100 m, 16 years old at 200 m and 400 m. In females, the correlation had a similar tendency to that of males and was strong until several years in the records of each school year, but gradually weakened after that. The correlation with the latest TIME became strong from 18 years old at 50 m, 16 years old at 100 m and 14 years old at 200 m and 1 to 2 years earlier than males. In the 400 m competition, there were few 12 years old subjects to analyze and it was not adapted for similar analysis.

The developmental changes in performance in the upper and lower groups
The developmental changes in performance of the upper and lower groups are shown in Figure-3 and -4 for male and female swimmers. The main effects of performance level and age factors were observed (p < 0.05) at all swimming distances for both males and females, but no interactions (performance level x age) were observed (Males

Discussion
This study is the first longitudinal analysis in Japan using a public database storing all records of official competitions accredited by the Japan Swimming Federation to analyze the process of how the performances changed from adolescence to adulthood.
In terms of performance maturity, female swimmers tend to develop faster, reaching their peak about 1 or 3 year earlier than their male counterparts (Figure-2). There was insignificant or no correlation between the malesʼ latest TIME and school-age TIME in any of 50 m, 100 m, 200 m, and 400 m. Females showed the same results in every swimming distance. The correlation between the latest TIME and school-age TIME became stronger in early adolescence or after 17 years old (Table-3 , 4). There was no diffrence in developmental changes in performance between the upper and lower groups at all swimming distances for both males and females (Figure-3, 4). Allen et al. 11) suggest that the age when swimmers show their peak performances comes earlier to females than males, and the age tends to be younger for both males and females as they swim longer distances. This research showed a similar trend that female swimmers tended to reach their peak earlier than their male counterparts.
Hollings et al. 12) states the age (mean ± SD) when track and field athletes show their peak performance as follows: males 25.2 ± 0.3 years old, females 25.7 ± 0.3 years old in short-distance race; males 24.9 ± 0.3 years old, females 26.7 ± 0.5 years old in middle-distance race. The age when swimmers reach their peak is younger than track and field athletes 11) . Swimmers show high performance maturity status of 80 to 90% during childhood, with a younger age to reach their peak than track and field athletes, which indicates the necessity of early swimming training in order to get top-class results at their peak age. However, the correlation analysis of this research shows no relationship between childhood and adulthood performance. These results are almost identical to the past report by Costa et al. 10) regarding the robustness of the swimming records using the public database of male elite swimmers in Portugal, which shows, as a result of the correlation analysis of the records of freestyle short-course evens in each age, no relationship between childhood performance and adulthood performance, and the relationship only becomes stronger after 15 or 16 years old. This indicates that high performance maturity during childhood is one of the required conditions for ultimately achieving top-class records, but the choice of training after childhood is the key factor of whether the athlete can achieve the top-class records. Identifying the age when the correlation with the latest TIME became stronger affects the choice of training for athletes during adolescence, or their growing stage. For malesʼ short-distance events such as 50 m and 100 m race, the correlation with the latest TIME tends to be strong during high school or later. As short-distance events require exerting a great force in a brief period of time, the fact that the records during when the athletesʼ strength and power significantly develop have a strong correlation with the latest TIME can relate to the trainability of strength and power during their growing stage. Ohsawa 13) estimated the appropriate age to start physical trainings for Japanese children from the relationship between their Peak Height Velocity (PHV) age and measurement data of physical fitness: PHV age is 11 years old for males, 9 years old for females; the best age to start physical (strength and power) trainings is 10.6 to 14.9 years old for males, 7.65 to 13.55 years old for females, earlier than males. The appropriate age to start building up endurance is estimated to be 10 years old for males and 9 years old for females, a year earlier than males. The research on male swimmers around their PHV age shows that the trainability of swimmersʼ strength and power development through resistance training responds better after PHV age than before PHV age 14) 15) . In addition, as a result of the resistance training given to swimmers before and after PHV age, improvements in strength and power can be seen after PHV age, while it was not work effective before PHV age 16) . The trainability of the strength and power is low in childhood and becomes higher from late adolescence to adulthood, when the height reaches its maximum, which can be related to the fact that the correlation between the achievements of short-distance events swimmers and their records in adulthood becomes stronger during 15 + or later.
On the other hand, for middle and long-distance events such as the 200 m and 400 m race, the correlation tends to be strong during junior-high school or high school, earlier than short-distance events 17) . The trainability during oneʼs growing stage is considered to affect their endurance capacity earlier than their strength and power. This explains why the correlation with the latest TIME can be seen in the middle and long-distance events, which require endurance capacity, earlier than the short-distance events. Furthermore, the PHV age, a benchmark of the acceleration point of the trainability of strength, power and endurance capacity, comes earlier to females than males, which can be related to the fact that the correlation between the femalesʼ latest TIME and school-age TIME in the middle and long-distance events can be confirmed earlier than males.
There was no difference in trend in developmental changes between the upper and lower group of swimmers ranked within the top 100 in adulthood in all swim distances and gender (Figure-3, 4). This means that strength, power and endurance trainability are related to the growing stage, not to swimming distance and gender. Therefore, training plans based on the growing stage of each swimmer are required to extend the swimmerʼs ability according to the characteristics of each swim distance (i.e., short-distance events require strength and power development, while middle and long-distance events require endurance capacity development).
In this research, the volume of data in each schoolage that meet the search criteria decreases as we date back to older years. The subjects of this study did not always participate in the same events since childhood. Changes to other swimming events or other sports could be the reason why there are not so many uninterrupted records since childhood. Sokolovas et al. 18) suggests that out of the top 100 elite swimmers in the U.S. aged from 17 to 18 years old, only 10% ranked in the top 100 between 11 to 12 years old, pointing out that the female elite swimmers often change their events by 13 to 14 years old, and male counterparts tend to change by 15 to 16 years old. Such situations might affect the results of this research. One of the reasons for athletes to change their event can be the influence of maturation rate. An athlete whose maturation rate is relatively fast shows good records compared to their counterparts of the same age not because of training, but because of early maturity. In such cases, setting appropriate goal is essential otherwise the athlete may later experience sluggish growth leading to drop out. On the contrary, an athlete whose maturation rate is slow relative to others may see a decrease in motivation due to a widening gap between rivals of the same age, which could cause a change of sporting event or drop out. Costa et al. reports on the relative age effect for swimmers in their youth and that the effect can be observed in 12-year-old females and 12-to 15year-old males 19) .

Limitations
This research used the public database of freestyle swimming competition of top-class swimmers from adolescent to adulthood to analyze the characteristics of the process of their performance improvements by tracing back through their past records. The possibility to analyze the growth process of Japanese top-class freestyle swimmers by using a public database in swimming competition was shown. However, this research has several limitations. First, we used only the record holders over 22 years old at the beginning of this analysis. Thus, not all swimmers in the top 100 were analyzed. Second, the extracted data was few at primary school time possibly affected by the change of competition style and the conversion of competition. Third, since this study only collected swimmersʼ performances, chronological age and anthropometric data were not used, and there is a limit to explain the relation to development. Fourth, because no swimming training content was collected, interpretation of the relationship between swimming training and performance must be considered.
In the future, in order to further clarify the process of swimmerʼs performance improvements, and obtain further results, it will be necessary to use the research design such as"follow-up study" to confirm the continuity of the swimming event from school age, "case-control study"with a control group, and"longitudinal study"linking physical ability and anthropometric data with performance. There is also an indication that gene polymorphisms such as angiotensin-converting enzyme (ACE) and α-actinin-3 (ACTN3) could affect the performances in competitive swimming of elite Caucasian and East-Asian swimmers and the events they choose. The relationship with these polymorphisms may also need to be studied 20) .

Conclusion
We used the official records of freestyle shortcourse events stored in the public database to analyze the swimmersʼ performances, through which we revealed the process of their performance improvements in line with their physical growth. The performance maturity of 12 years old is above 80% for males and 90% for females, indicating that the swimmers ranked as top-class during adulthood have already shown superior performances in childhood. On the other hand, there is no or insignificant correlation between the records of adulthood and childhood, only becoming strong after adolescence, which suggests the records of childhood is a necessary condition for discovering promising swimmers who could achieve top-class results in adulthood, but not a sufficient condition. The relationship with the records of adulthood tends to become strong in the late adolescent period for sprint events which mainly require strength and power, and in early adolescence for middle and long-distance events which mainly require endurance. Therefore, taking the swimmerʼs maturity into consideration is necessary for properly choosing trainings and setting goals.